A MODIFIED POWELL METHOD FOR 6-DEGREES-OF-FREEDOM PLATFORM KINEMATICS

被引:6
作者
DESHMUKH, G [1 ]
PECHT, M [1 ]
机构
[1] UNIV MARYLAND,COLLEGE PK,MD 20742
关键词
D O I
10.1016/0045-7949(90)90273-5
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
As an alternative to using the Newton-Raphson method for determining inverse transformations for the real-time simulation of a six-degrees-of-freedom platform, a modified Powell method was developed which couples the steepest descent technique, Broyden's method for approximating the Jacobian matrix and its inverse, and Marquardt's method to minimize the sum of squares of the functions. It was found that the modified Powell method for random initial values is superior to the Newton-Raphson, Broyden, and Marquardt methods with respect to the number of function evaluations and computational time. © 1990.
引用
收藏
页码:485 / 491
页数:7
相关论文
共 8 条
[1]  
BROYDEN CG, 1965, MATH COMPUT, V19, P577, DOI DOI 10.1090/S0025-5718-1965-0198670-6
[2]  
DIEUDONNE JE, 1972, D7067 NAT AER SPAC A
[3]  
HASELGROVE CB, 1961, COMPUT JNL, P255
[4]  
LEVENBERG K, 1943, NOTES ANN M AM MATH, V2, P164
[5]   AN ALGORITHM FOR LEAST-SQUARES ESTIMATION OF NONLINEAR PARAMETERS [J].
MARQUARDT, DW .
JOURNAL OF THE SOCIETY FOR INDUSTRIAL AND APPLIED MATHEMATICS, 1963, 11 (02) :431-441
[6]  
Paul R. P., 1981, ROBOT MANIPULATORS M
[7]  
POWELL MJD, 1968, AERE5947 MATH BRANCH
[8]  
PREMACK T, 1984, NASA86111 TECHN MEM